RE: Cryptogram Generator

[btrott]

A couple of comments:

1) the open || die statement should actually be

open IN, $infile or die "Can't open input file: $!";

For reasons of precedence.

2) Here are some faster algorithms. One way is to use a hash to record which characters you've already placed into your substitution string. Here's the main loop:

my %set;
my $substit = "";
for (1..26) {
    my $randchar;
    do { $randchar = chr((int rand 26) + 65) }
        while $set{$randchar}++;
    $substit .= $randchar;
}

This way you don't have to do a search through the string each time.

And here's an even better way of doing it: build a random permutation of the alphabet string. To do this, we'll actually need an array:

my @alpha = split //, "ABCDEFGHIJKLMNOPQRSTUVWXYZ";

Now we set another array equal to this one, then call the random shuffle algorithm on it:

my @crypt = @alpha;
fisher_yates_shuffle(\@crypt);

which shuffles the array in place. Now all you need to do is get back the substitution string:

my $substit = join '', @crypt;

Here's the definition of the fisher_yates_shuffle sub:

sub fisher_yates_shuffle {
    my $array = shift;
    for (my $i = @$array; --$i; ) {
        my $j = int rand ($i+1);
        next if $i == $j;
        @$array[$i, $j] = @$array[$j, $i];
    }
    return join '', @$array;
}

(Taken directly from perlfaq4.)

I did some benchmarking on these, and here's what I got:

Benchmark: timing 5000 iterations of orig, f_yates, hash...
      orig: 33 secs (23.60 usr  0.00 sys = 23.60 cpu)
   f_yates:  5 secs ( 2.88 usr  0.00 sys =  2.88 cpu)
      hash: 13 secs ( 5.68 usr  0.00 sys =  5.68 cpu)

where "orig" is the one you posted, "f_yates" is the one using fisher_yates_shuffle, and "hash" is the one using a hash to record seen characters.

If I've messed anything up, let me know.

[Elihu]

Wow. Thanks for the new ideas. I've been playing with the different algorithms a bit. I've seen somewhere how you do benchmarking, I'll try my own just to see for myself how it's done and if I get comparable numbers.

I was pretty sure there was a better way to do that. :)

[duggles]

I've been working on a cryptogram solver and discovered this post and found the code useful - I snagged the code for the fisher_yates_shuffle to grab one of my solved cryptograms to recreate one to work on solving and I found that in some instances one or two of the letters in $substit remained the same as it was originally. The line that checks for $i == $j does a next BUT $i keeps on counting down and effectively leaves the unchanged letter alone. I fixed it by incrementing $i before doing the next.

I realize in some cases the shuffle results are okay if this happens, but in a cryptogram, it won't fly... ;-) Just thought I'd throw this info back out there just in case you or anyone else may have a need for the fix.

Here's the modified sub:

## Taken from perlfaq4 (thanks btrott)
sub fisher_yates_shuffle {
  my $array = shift;
  for (my $i = @$array; --$i; ) {    
    my $j = int rand ($i+1);
    # next if $i == $j; # original
    if ($i == $j) {  # this means the letter will be the same as it wa
+s before 
        $i++; # put $i back where it was and get another one
        next; 
    } 
    @$array[$i, $j] = @$array[$j, $i];
  }
  return join '', @$array;
}

thanks for the original post! I'm learning a little here and there about using perl to solve cryptos... it's gonna take a while, but it's fun!!!